The research is designed to test hypotheses about the influence of specific personality variables on level of alcohol consumption in middle-aged and young adult females and males and on changes in consumption in young adults over a three-year period. The investigative effort focuses on three variables; dependency conflict, allegiance conflict and sex-role conflict. Tests of the effect of these three types of conflict on alcohol consumption will be pursued through analyses of existing data in the Family Dynamics Study data base (Schwarz, 1983). This data base, constructed over the last six years, consists of scores on over 75 assessment instruments obtained from (cir) 1,800 individuals associated with 360 family units. Each unit consists of an adolescent college student (the primary subject), the student's roommate, mother, father and a sibling. The students (190 female and 170 male) were about 18 years of age at the initial assessment and about 21 years at the follow-up. Included are extensive data on the subjects' sex-role behavior, sibling-position status, attitudes, personality, psychopathologic symptoms, alcohol and other drug use, and relationships with parents. In addition, the data base contains data on about 330 mothers and about 320 fathers with respect to their alcohol consumption; marital interaction; childrearing behaviors; personality dimensions; competence, adjustment; sexual, social and political attitudes; employment history; and sibling-position status. It is assumed that alcohol reduces fear and that individuals who experience approach-avoidance conflict will find alcohol consumption highly reinforcing and, therefore, be at greater risk for alcohol abuse and dependency. The dependency-conflict hypothesis (which is consistent with higher alcoholism among latter borns from large families) will be tested using parental childrearing behaviors and sibling-position data to assess dependency conflict in concurrent and prospective prediction of alcohol consumption. The allegiance-conflict hypothesis will be tested by comparing levels of alcohol consumption between young adults from harmonious and maritally discordant homes in a complex design that also assesses differential parental modeling effects. Extremely high rates of alcohol abuse in both male and female homosexuals suggest that sex role conflict may be predisposing toward alcohol abuse. A new theory of sex-role development is proposed which promises to yield more precise measurement of sex-role conflict, and therefore better prediction of alcohol abuse and dependency, should this hypothesis be supported.